Time, Technology, Talent: the Three- Pronged Promise of Cloud ML Cloud ML Appears to Be the Future of AI, and the Future Seems Bright
Total Page:16
File Type:pdf, Size:1020Kb
FEATURE Time, technology, talent: The three- pronged promise of cloud ML Cloud ML appears to be the future of AI, and the future seems bright Sudi Bhattacharya, Ashwin Patil, Jonathan Holdowsky, and Diana Kearns-Manolatos THE DELOITTE CENTER FOR INTEGRATED RESEARCH Time, technology, talent: The three-pronged promise of cloud ML Start with the business case, assess your data and model requirements, invest in the right talent, and be willing to fail fast to reap the maximum benefits from cloud machine learning. S ORGANIZATIONS LOOK to use machine services, and platforms—including pretrained learning (ML) to enhance their business models and accelerators—that provide more Astrategies and operations, the retailer options for how data science and engineering Wayfair looked to the cloud. Wayfair believes what teams can collaborate to bring models from the lab so many other organizations seem to—that the to the enterprise. Indeed, because of all these cloud and the fully integrated suite of services it advantages, time, technology, and talent—which offers may give organizations the best approach to might otherwise serve as challenges to scaling create ML solutions with time to market, existing enterprise technology—comprise the three- resources, and available technologies. For its part, pronged promise of cloud ML (see sidebar, “The Wayfair pursued the cloud for its scalability—key measurable value of cloud ML”). for retailers with varying demand levels—as well as a full suite of integrated analytics and productivity The recognition of this potential has fueled rapid tools service offerings to drive actionable insights. growth: The current cloud ML market is estimated The results? Faster insights into changing market to be worth between US$2 billion and US$5 billion, conditions to save time. Immediate access to the with the potential to reach US$13 billion by 2025. most current ML technologies. And tools of What such growth says is that adopters are productivity that can drive efficient talent.1 increasingly realizing the benefits of cloud ML and that it is the future of artificial intelligence (AI). Indeed, ML is transforming nearly every industry, And with just 5% of the potential cloud ML market helping companies drive efficiencies, enable rapid being penetrated according to one estimate, the innovation, and meet customer needs at an community of adopters should only deepen across unprecedented pace and scale. As companies sectors and applications.2 increasingly mature in their use of ML, the heavy reliance on large data sets and the need for fast, To explore the potential cloud ML represents for reliable processing power are expected to drive the organizations looking to innovate, we conducted future of ML into the cloud. interviews with nine leaders in cloud and AI whose perspectives largely inform the content that follows. The well-known benefits of the cloud—including In this paper, we provide an overview of three modular, elastic, rapidly deployable, and scalable distinct operating models for cloud ML and offer infrastructure without heavy upfront investment— insights for executives as they seek to unlock the apply when bringing ML into the cloud. Cloud ML full potential of the technology. additionally introduces cutting-edge technologies, 2 Cloud ML appears to be the future of AI, and the future seems bright THE MEASURABLE VALUE OF CLOUD ML In the 2020 Deloitte State of AI in the Enterprise survey, 83% of respondents said that AI will be critically or very important to their organizations’ success within the next two years. Around two- thirds of respondents said they currently use ML as part of their AI initiatives. And our survey suggests that a majority of these AI/ML programs currently use or plan to use cloud infrastructure in some form. Our additional analysis of the survey data for this research showed that cloud ML, specifically, drives measurable benefits for the AI program (figure 1) and generally improves outcomes as compared to nonspecific AI deployments: • Forty-nine percent cloud ML users said they saw “highly” improved process efficiencies (vs. 42% overall survey results) • Forty-five percent saw “highly” improved decision-making (vs. 39% overall) • Thirty-nine percent experienced “significant” competitive advantage (vs. 26% overall) Importantly, these data points showed that cloud ML tools appear to make a notable impact on longer-term, strategic areas, such as competitive advantage and improved decision-making, when compared to general AI adoption. This is an important distinction and shift in the approach organizations are taking to run their AI programs (away from on-premise AI platforms), and a trend that is expected to continue as cloud ML takes greater focus for global organizations. The addition of low-code/no-code AI tools, which are part of some cloud ML service offerings, can extend these advantages with responses showing additional gains in decision-making, process efficiencies, and competitive advantage and other areas. For instance: • Fifty-six percent of low-code/no-code cloud ML users said they saw “highly” improved process efficiencies (vs. 49% in cloud ML users vs. 42% overall) • Fifty-four percent saw “highly” improved decision-making (vs. 45% in cloud ML users vs. 39% overall) • Forty-nine percent experienced “significant” competitive advantage (vs. 39% in cloud ML users vs. 26% overall) • Low-code/no-code cloud ML users saw additional gains in areas that general cloud ML users did not see with outcomes such as new insights and improved employee productivity, among others. The emergence of low-code/no-code tools is part of a broader trend toward greater accessibility to cloud AI/ML solutions and a more diverse user community. These positive survey results suggest that this trend may only strengthen in the near to medium term. 3 Time, technology, talent: The three-pronged promise of cloud ML Cloud ML can drive improved outcomes in key areas of decision-making, process efficiencies, and competitive advantage Total surey respondents loud-ased as a serice loud-ased ML loud-ased ML ith lo-codeno-code Percentage who achieved outcome to a “high” degree Lowering cost 33% 32% 30% 29% Reducing headcount 31% 31% 30% 26% Improving decision-making 39% 42% 45% 54% Making processes more efficient 42% 46% 49% 56% Enhancing relationships with clients/ customers 36% 37% 38% 41% Making employees more productive 36% 38% 40% 49% Discovering new insights 36% 38% 40% 54% Enhancing existing products and services 37% 37% 37% 38% Creating new products and services 35% 36% 38% 40% Enabling new business models 34% 33% 32% 40% Enabling a “significant” lead over competitors 26% 34% 39% 49% ote: s a prereuisite to participate, all , respondents to the surey must hae adopted technologies in at least some fashion, ased on any of seeral criteria ee surey methodology for more detailed description of criteria ources: eena mmanath, aid aris, and usanne upfer, Thriing in the era of perasie : eloittes tate of in the nterprise, rd dition, eloitte nsights, uly , eloitte analysis Deloitte Insights | deloitte.com/insights 4 Cloud ML appears to be the future of AI, and the future seems bright Cloud ML: Adopting the right the organization is trying to solve, the technology approach context (to what degree the organization has data and models on-premises or in the cloud already), “AI is a tool to help solve a problem,” says Rajen and the talent context (to what extent the Sheth, vice president, AI, Google Cloud. “[In the organization has the right people resources in place past,] people were doing a lot of cool things with AI to build and train models). that didn’t solve an urgent problem. As a result, their projects stopped at the concept stage. We now There is no one-size-fits-all approach. Depending start with a business problem and then figure out on the business problem, multiple approaches may how AI solves it to create real value.” be employed at the same company. As Gordon Heinrich, senior solutions architect at Amazon Web As organizations look to advance their ML Services, says, “At the end of the day, it comes down programs supported by the cloud, there are three to where you can put ML and AI into a business basic approaches to cloud ML (figure 2) they can process that yields a better result than you are take—cloud AI platforms (model management in currently doing.” Building a model in-house the cloud); cloud ML services (including pretrained requires significant investment of time in data models); and AutoML (off-the-shelf models trained collection, feature engineering, and model with proprietary data). development. Depending on the specific use case and its strategic importance, companies can employ Selecting the right approach starts with cloud ML services, including pretrained models, to understanding the business context—what problem help speed up time to innovation and deliver results. FIGURE 2 Cloud ML: Three core approaches and their benefits Cloud AI platforms: Model management Cloud ML services: Pretrained AutoML: Customized in the cloud models and more pretrained models What it is Brings new and Allows organizations to access Allows for “customizing” off-the- existing ML models pretrained models, frameworks, shelf versions of pretrained API into the cloud and general-purpose algorithms models with proprietary data including: Allows developers • Vision, speech, language, and Allows for ongoing model and data scientists video APIs refinement, which, in some cases, to use cloud-enabled • Virtual assistant